I just discovered (via Tyler Cowen) a fascinating economics paper on the changing dynamics of scientific production over the 20th century. A few months ago, I wrote about the tangled relationship of age and innovation, and why different fields have different peak ages of creativity. In general, math, physics and poetry are for the young, while biology, history and the social sciences benefit from middle-age:
Interestingly, these differences in peak age appear to be cultural universals, with poets peaking before novelists in every major literary tradition, according to [Dean] Simonton's research.
What accounts for these variations? Mr. Simonton suggests that they're caused by intrinsic features of the disciplines. Those fields with a logically consistent set of principles, such as physics and chess, tend to encourage youthful productivity, since it's relatively easy to acquire the necessary expertise. (The No. 1 ranked chess player in the world today, Magnus Carlsen, is 19 years old.) Because the essential facts can be quickly learned, and it usually doesn't take that long to write a lyric poem, the precocious student is free to begin innovating at an early age.
In contrast, fields that are loosely defined and full of ambiguous concepts, such as biology and history, lead to later peak productive ages. After all, before a researcher can invent a useful new idea, he or she must first learn an intimidating assortment of details.
This new economics paper, by Benjamin Jones at Northwestern, describes a fascinating natural experiment that captures this dynamic at work:
Jones and Weinberg further analyze the effect of an exogenous shock to the foundational knowledge in a field, studying the age and training patterns around the quantum mechanics revolution in physics. The quantum mechanics revolution is typically charted between 1900 and 1927. Remarkably, we find that (a) age at great achievement and (b) age at Ph.D. actually declined in physics during this period, reaching a minimum just as quantum mechanics becomes a rigorously established theory in the late 1920s and then rising thereafter. Moreover, these patterns are unique to physics; the age of great achievements and Ph.D. age in other fields continued to rise during this period. Viewed as a natural experiment, the analysis of the quantum mechanics revolution further substantiates the link between the current depth of knowledge in a field, its training requirements, and the ensuing innovative output of young scholars.
Why were young physicists better at working on quantum mechanics, at least in the early part of the 20th century? One possibility is that the youthful brain is intrinsically more creative, and thus better able to contemplate the surreal properties of subatomic particles. In other words, the creativity of youth is really a story about cognitive decline and the inevitable atrophy of the human cortex. But that's almost certainly not the case. (After all, some academic fields, such as literary criticism, have a peak creative age in the late forties.) Instead, Simonton and others argue that young physicists benefit, at least in part, from their outsider status - they're more innocent and ignorant - which makes them more willing to embrace novelty and surprise. Because they haven't become "encultured," or weighted down with too much conventional wisdom, they're more likely to rebel against the status-quo and explore the the spooky ideas of Schrodinger, Bohr, et. al. After a few years in the academy, however, Simonton says that "creators start to repeat themselves, so that it becomes more of the same-old, same-old." They have become insiders, invested in Newtonian mechanics; that is what they know and that is what they believe in. It's only the impetuous youth, those marginal figures without tenure or grants of their own, who properly appreciate the anomalies of the subatomic world.
What makes this "natural experiment" so convincing is that it counters the larger trend of 20th century science, in which the peak age of scientists got older and older. (According to Jones, the peak age was extended by approximately five years during the 20th century.) This is largely a side-effect of scientific success: researchers discovered a lot of new facts, which meant that it took a few additional years to master any given field.
I'd be interested in seeing if this same pattern holds true for every paradigm shift/scientific revolution. Did the rise of molecular biology lead to a decrease in the peak age of molecular biologists? What about the modern synthesis in evolutionary biology? Or the rise of the modernist novel? Alternatively, it's also possible that the quantum revolution was a relatively unique scientific moment, an "exogenous shock" that was unusually shocking.
I'm not that familiar with the history of QM beyond the canned version they teach in physics classes, but wouldn't there have been a large amount of low hanging fruit in the early days after QM was established?
To get an equation named after you, you just had to be first to apply QM to a given system or problem, and some of these would be easier than others. After a while, the "easy" stuff was done and so it took longer to really make a mark in the field.
Other sciences may have had similar situations occur, but probably not to the same extreme degree.
I buy it. Heisenberg famously challenged Bohr during a lecture when Heisenberg was just 20 years old, and so impressed Bohr that he took Heisenberg on as a student. And Einstein's most productive year came at the young age of 26!
But it's just inconceivable to imagine these things happening today. There's simply so much more physics and mathematics that one needs to know nowadays to make significant contributions in fundamental physics.
A student interested in fundamental theory today needs to master statistical mechanics, quantum mechanics, quantum field theory, and general relativity, all at a post-graduate level, before anyone would take him or her seriously. None of these subjects (except statistical mechanics) existed as an established, broadly-known subject back in the 1920s. And it's just impossible to squeeze all of that down to a few years of study. It takes a very, very long time to get up to speed in the contemporary era.
And I haven't even mentioned the advanced mathematics that a physics student interested in fundamental physics needs to know these days, ranging from differential geometry and complex analysis to group theory, let alone esoteric subjects like string theory and algebraic geometry whose methods have spilled out into particle phenomenology in recent years.
It's therefore little surprise that most of the breakthroughs in fundamental theoretical physics have come from older physicists; the discovery of M-theory and the web of string-theory dualities by Witten, D-branes by Polchinski, the AdS/CFT correspondence by Maldacena, black-hole entropy-counting formulas by Strominger and Vafa, brane-world models by Randall, etc. None of these physicists were very young when they made these breakthroughs. And the age situation is even more striking for experimental particle physicists.
Fundamental physics has unavoidably gotten much more like biology and history in recent years, at least in terms of the vast learning curve faced by its would-be practitioners. Nature was very kind to us in the early 20th century, but has become much more stingy in recent times.
My hope is that some of this problem might be mitigated by revisiting the way we teach these subjects; they've become weighed down by the contingencies of history, and syllabi could probably be compressed, unified, and streamlined significantly if we approached them with fresh eyes and hindsight. But I still think the situation is largely irreversible.
Typo: "It's therefore little surprise that most of the relatively recent breakthroughs in fundamental theoretical physics have come from older physicists."
(By the way, Jonah, I adore your recent sudden interest in all things quantum mechanics!)
"Viewed as a natural experiment, the analysis of the quantum mechanics revolution further substantiates the link between the current depth of knowledge in a field, its training requirements, and the ensuing innovative output of young scholars."
Errr, actually it demonstrates the exact opposite, if your claim that the minimum age corresponded to the END of the period of rapid expansion in information needing to be absorbed happens to be accurate. For it to demonstrate what you claim, the age would have to INCREASE most rapidly during periods of rapid uncovering of new knowledge, and more slowly at other times. And indeed, never to decrease at all, except at the start of the Dark Ages :-)
The Blunderbuss effect is quite sufficient to explain a sudden crop of youthful apparent geniuses in a new field. You simply don't yet know that most of them are wrong.
"Why were young physicists better at working on quantum mechanics"
You don't know that they were. To even hazard a guess that that you would ALSO have needed to collect data on rejected papers, papers which were accepted but subsequently rejected, and failed PhDs. If they showed an even greater preponderance of youngsters, then your claim simply wouldn't hold water.
My brain just struck some analogous chord with rock music. All the great stuff was done in the first 25 years or so after Elvis and the rest has mostly been nuanced from those early influences. Don't get me wrong, there is still great song writing going on today. Isn't that a lot like QM? Although getting the Higgs boson squeezed out of the Hadron collider could do some paradigm smashing. Reminds me of that ol' Sam Cooke song: "Don't know much history, don't know much biology...."
I'm a graduate student in the history of science, and one of my undergraduate degrees was chemistry, so I recognize the importance youth has had in many disciplines when they were just getting started. Now, however, so many graduate programs require literature reviews in dissertations, etc., and scholars have produced so much in the last century that to be young and innovative is almost impossible. The sciences are a little more flexible in this regard, but in history, 90% of my time has been devoted to reading other scholars, as opposed to doing my own research. This could be a problem with the field in general, or maybe it's that theory is supposedly what lazy scholars did in the 1960s when they couldn't be bothered to go to the archives themselves. Today, you can't get published without a degree, and you can't get a degree without understanding/catering to established systems of thought. There are reasons so many academics go crazy.
Re comments 6 & 7: Rock music had great stuff? Perhaps an alternative explanation is that many young performers either failed ("one-hit wonders"), made lots of money and retired, or did lots of drugs and OD'd/burned out. (Or perhaps they simply developed better taste as they aged :-)) To a casual observer like me, it certainly seems that there's a great deal of attrition in the field, so the reason that there's little great stuff (however defined) from older rock musicians is that there just aren't all that many old rockers.
But for an alternative, look at the classical composers: Bach, Beethoven, Hayden & others were all composing creative works (and arguably equaling or bettering their earlier ones) until shortly before their deathis in their 50s, and later.
Jonah, I think that the age levels are based on correlation, rather than causation.
I have read elsewhere that peak neural growth happens at ages 17 and 21.
This study seems relevant - http://www.springerlink.com/content/a2j2nq583tgwj99x/
I think this is not based in the mind. I am too much of a Chomskian to accept that radical of a behaviorist view.
Rather, I think that, (from a student's perspective), rapid growth happens when the brain is comfortable with itself. Senior year in high school, and also in college (roughly ages 17 and 21) are the years when the mind is able to have fun.
Fun and learning ALWAYS go hand in hand.
Have you read anything about super-learning?
I have posted on this before, though my writing isn't nearly as clear as yours... I'm still a blogging newbie
So many things differ about that period. There were more physicists at Berkeley in 1975 than in the WORLD in 1932, about 300, by which time most of the theoretical discoveries had been made. Physics wasn't much of a career. Einstein wasn't even employed as a physicist in 1905!
Great Article Jonah.
This is why the Karl Popper had reminded us that the Gold standard for seeking truth is not trying to prove but to refute what you are trying to actually prove.